Hydrocarbons Are Essential for Optimal Cell Size, · a role for hydrocarbons in limiting cell size...

Hydrocarbons Are Essential for Optimal Cell Size,Division, and Growth of Cyanobacteria1[OPEN]

David J. Lea-Smith*, Maite L. Ortiz-Suarez, Tchern Lenn, Dennis J. Nürnberg2, Laura L. Baers,Matthew P. Davey, Lucia Parolini, Roland G. Huber, Charles A.R. Cotton2, Giulia Mastroianni,Paolo Bombelli, Petra Ungerer, Tim J. Stevens, Alison G. Smith, Peter J. Bond, Conrad W. Mullineaux,and Christopher J. Howe

Department of Biochemistry, University of Cambridge, Cambridge CB2 1QW, United Kingdom (D.J.L.-S.,L.L.B., C.A.R.C., P.B., C.J.H.); Centre for Molecular Science Informatics, Department of Chemistry, Universityof Cambridge, Cambridge CB2 1EW, United Kingdom (M.L.O.-S., P.J.B.); School of Biological and ChemicalSciences, Queen Mary University of London, London E1 4NS, United Kingdom (T.L., D.J.N., G.M., P.U.,C.W.M.); Department of Plant Sciences, University of Cambridge, Cambridge CB2 3EA, United Kingdom(M.P.D., A.G.S.); Department of Physics, University of Cambridge, Cambridge CB3 0HE, United Kingdom(L.P.); Bioinformatics Institute, A*STAR, Singapore 138671 (R.G.H., P.J.B.); MRC Laboratory of MolecularBiology, Cambridge CB2 0QH, United Kingdom (T.J.S.); and National University of Singapore, Departmentof Biological Sciences, Singapore 117543 (P.J.B.)

ORCID IDs: 0000-0002-2980-9749 (M.L.O.-S.); 0000-0002-5827-4616 (T.L.); 0000-0001-6511-5704 (A.G.S.); 0000-0003-2900-098X (P.J.B.);0000-0001-7194-9916 (C.W.M.).

Cyanobacteria are intricately organized, incorporating an array of internal thylakoid membranes, the site of photosynthesis, intocells no larger than other bacteria. They also synthesize C15-C19 alkanes and alkenes, which results in substantial production ofhydrocarbons in the environment. All sequenced cyanobacteria encode hydrocarbon biosynthesis pathways, suggesting animportant, undefined physiological role for these compounds. Here, we demonstrate that hydrocarbon-deficient mutants ofSynechococcus sp. PCC 7002 and Synechocystis sp. PCC 6803 exhibit significant phenotypic differences from wild type, includingenlarged cell size, reduced growth, and increased division defects. Photosynthetic rates were similar between strains, although aminor reduction in energy transfer between the soluble light harvesting phycobilisome complex and membrane-bound photosystemswas observed. Hydrocarbons were shown to accumulate in thylakoid and cytoplasmic membranes. Modeling of membranessuggests these compounds aggregate in the center of the lipid bilayer, potentially promoting membrane flexibility andfacilitating curvature. In vivo measurements confirmed that Synechococcus sp. PCC 7002 mutants lacking hydrocarbonsexhibit reduced thylakoid membrane curvature compared to wild type. We propose that hydrocarbons may have a role ininducing the flexibility in membranes required for optimal cell division, size, and growth, and efficient association of solubleand membrane bound proteins. The recent identification of C15-C17 alkanes and alkenes in microalgal species suggestshydrocarbons may serve a similar function in a broad range of photosynthetic organisms.

Cyanobacteria (oxygenic photosynthetic bacteria) arefound in nearly every environment on Earth and aremajor contributors to global carbon and nitrogen fixa-tion (Galloway et al., 2004; Zwirglmaier et al., 2008).They are distinguished among prokaryotes in contain-ing multiple internal thylakoid membranes, the site ofphotosynthesis, and a large protein compartment, thecarboxysome, involved in carbon fixation. Despite theseextra features, cyanobacteria can be as small as 0.6 mmin diameter (Raven, 1998).

All cyanobacteria with sequenced genomes encodethe pathway for the biosynthesis of hydrocarbons, im-plying an important, although as-yet-undefined, rolefor these compounds (Lea-Smith et al., 2015). The majorforms are C15-C19 alkanes and alkenes, which canbe synthesized from fatty acyl-acyl-carrier proteins(ACPs) by one or other of two separate pathways (Fig.1; Schirmer et al., 2010; Mendez-Perez et al., 2011). Themajority of species produce alkanes and alkenes via

acyl-ACP reductase (FAR) and aldehyde deformylatingoxygenase (FAD; Schirmer et al., 2010; Li et al., 2012;Coates et al., 2014; Lea-Smith et al., 2015). Cyano-bacterial species lacking the FAR/FAD pathway syn-thesize alkenes via olefin synthase (Ols; Mendez-Perezet al., 2011; Coates et al., 2014; Lea-Smith et al., 2015).This suggests that hydrocarbons produced by eitherpathway serve a similar role in the cell. Homologs ofFAR/FAD or Ols are not present in other bacteria orplant and algal species. However, C15-C17 alkanes andalkenes, synthesized by an alternate, uncharacterizedpathway, were recently detected in a range of greenmicroalgae, including Chlamydomonas reinhardtii,Chlorella variabilis NC64A, and several Nannochloropsisspecies (Sorigué et al., 2016). In C. reinhardtii, hydro-carbons were primarily localized to the chloroplast,which originated in evolution from a cyanobacteriumthat was engulfed by a host organism (Howe et al.,2008). Hydrocarbons may therefore have a similar role

1928 Plant Physiology�, November 2016, Vol. 172, pp. 1928–1940, www.plantphysiol.org � 2016 American Society of Plant Biologists. All Rights Reserved. www.plantphysiol.orgon March 14, 2020 - Published by Downloaded from

Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://orcid.org/0000-0002-2980-9749

http://orcid.org/0000-0002-5827-4616

http://orcid.org/0000-0001-6511-5704

http://orcid.org/0000-0003-2900-098X

http://orcid.org/0000-0001-7194-9916

http://www.plantphysiol.org

in cyanobacteria, some green microalgae species, andpossibly a broader range of photosynthetic organisms.Hydrocarbons act as antidesiccants, waterproofing

agents, and signalingmolecules in insects (Howard andBlomquist, 2005) and prevent water loss, ensure pollenviability, and influence pathogen interactions in plants(Kosma et al., 2009; Bourdenx et al., 2011). However, thefunction of hydrocarbons in cyanobacteria has not beendetermined. Characterization of cyanobacterial hydro-carbon biosynthesis pathways has provided the basisfor investigating synthetic microbial biofuel systems,which may be a renewable substitute for fossil fuels(Schirmer et al., 2010; Choi and Lee, 2013; Howard et al.,2013). However, secretion of long-chain hydrocarbons

from the cell into the medium, which is likely essentialfor commercially viable production, has not been ob-served in the absence of a membrane solubilizationagent (Schirmer et al., 2010; Tan et al., 2011). Cyano-bacterial hydrocarbons also have a significant envi-ronmental role. Due to the abundance of cyanobacteriain the environment, hydrocarbon production is consid-erable, with hundreds of millions of tons released intothe ocean per annum following cell death (Lea-Smithet al., 2015). This production may be sufficient to sus-tain populations of hydrocarbon-degrading bacteria,which can then play an important role in consuminganthropogenic oil spills (Lea-Smith et al., 2015).

Here, we investigated the cellular location and role ofhydrocarbons in both spherical Synechocystis sp. PCC6803 (Synechocystis) and rod-shaped Synechococcus sp.PCC 7002 (Synechococcus) cells. We developed a modelof the cyanobacterial membrane, which indicated thathydrocarbons aggregate in the middle of the lipid bi-layer and, when present at levels observed in cells, leadto membrane swelling associated with pools of hydro-carbon. This suggested that alkanes may facilitate mem-brane curvature. In vivo measurements of Synechococcusthylakoid membrane conformation are consistent withthis model.

RESULTS

Hydrocarbons Predominantly Localize to Thylakoid andCytoplasmic Membranes

Recently we demonstrated that 115 sequenced cya-nobacteria isolated from a broad range of environmentscontain either the far/fad or ols genes, encoding the en-zymes for alkane/alkene biosynthesis (Lea-Smith et al.,2015). In an additional 32 recently sequenced genomesfrom cyanobacteria, we found the same situation withthe majority, 133/147, containing far/fad homologs(Supplemental Table S1). Clearly, there is an importantrole for these compounds in cyanobacteria. In order toinvestigate this, we disrupted the two different biosyn-thetic pathways in two species of cyanobacteria thatare also morphologically distinct. Far in Synechocystisand ols in Synechococcus were disrupted by insertion ofa kanamycin resistance cassette into the open readingframe (Supplemental Fig. S1). In wild-type Synechocystis,1.44 mg/g dry cell weight of alkanes, predominantlyheptadecane and 8-heptadecene,were detected (Tan et al.,2011), whereas in Synechococcus 0.61mg/g dry cell weightof alkenes, specifically nonadecene (Mendez-Perez et al.,2011), were present (Supplemental Fig. S2). In contrast, inmutant cells lacking either FAR or Ols, no hydrocarbonswere observed. Complementation ofDFARby insertion offar into a neutral site on the chromosome restored alkanesto wild-type levels (Supplemental Figs. S1 and S2).

Due to their hydrophobic and nonpolar characteristics,hydrocarbons were expected to localize predominantly tomembranes. This was confirmed in purified plasma andthylakoid membrane fractions from Synechocystis (Fig. 2,A and B). Alkanes constituted 5.63% and 17.41% of the

Figure 1. Hydrocarbon biosynthesis is encoded in all sequenced cya-nobacteria. Detailed are the two hydrocarbon biosynthetic pathways,indicated in blue and red, respectively, in cyanobacteria. The number ofspecies encoding the enzymes in each pathway is indicated.

1 T.L. was supported by Biotechnology and Biological SciencesResearch Council (BBSRC) Research Grant BB/J016985/1 toC.W.M. D.J.L.-S. was supported by the Environmental Services As-sociation Education Trust. L.L.B. was supported by a BBSRCDoctoralTraining Grant (BB/F017464/1).

2 Present address: Department of Life Sciences, Imperial CollegeLondon, London SW7 2AZ, UK.

* Address correspondence to [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:David J. Lea-Smith ([email protected]).

D.J.L.-S. and C.J.H. conceived the original screening and researchplans; D.J.L.-S., P.J.B., C.W.M., and C.J.H. supervised the experi-ments; D.J.L.-S., M.L.O.-S., T.L., D.J.N., L.L.B., M.P.D., L.P., R.G.H.,C.A.R.C., G.M., P.B., and P.U. performed most of the experiments;D.J.L.-S., M.L.O.-S., T.L., D.J.N., L.L.B., M.P.D., R.G.H., T.J.S., P.J.B.,C.W.M., and C.J.H. designed the experiments and analyzed the data;D.J.L.-S. conceived the project and wrote the article with contribu-tions of all the authors; D.J.L.-S., P.J.B., C.W.M., and C.J.H. supervisedand complemented the writing.

[OPEN] Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.16.01205

Plant Physiol. Vol. 172, 2016 1929

The Role of Hydrocarbons in Cyanobacteria

www.plantphysiol.orgon March 14, 2020 - Published by Downloaded from Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org/cgi/content/full/pp.16.01205/DC1





mailto:[email protected]


mailto:[email protected]

http://www.plantphysiol.org/cgi/doi/10.1104/pp.16.01205


plasma and thylakoid membrane lipid fractions, respec-tively (Fig. 2C; Supplemental Fig. S3). Alkanes comprised8.92% of the total Synechocystis membrane lipid fraction.Given that thylakoids constitute a larger proportion ofcellular membranes than plasma membranes, this sug-gests that a hydrocarbon-rich portion of the thylakoidmembrane was purified during this process. In totalSynechococcus membrane fractions, alkenes consti-tuted 5.34% of total lipids (Fig. 2C).

Hydrocarbon-Deficient Strains Exhibit Enlarged Cell Sizeand Division Defects

To determine how loss of hydrocarbons affects cellmorphology,we used bright-fieldmicroscopy.DFAR cellswere significantly larger than wild-type Synechocystis(11.02 versus 4.63 mm3; Fig. 3, A and B; Supplemental Fig.S4; Supplemental Table S2), which was confirmed viaparticle-counting measurements (11.49 versus 4.58 mm3;Fig. 3C; Supplemental Table S3). In addition, a signifi-cantly larger percentage of DFAR cells were actively di-viding (47.4% versus 40.1%; Fig. 3F; Supplemental TableS4). Division defects were also apparent in DOls, whichformed long chains of up to 12 cells and abnormal rods(Fig. 3D; Supplemental Fig. S4). The width of DOls cellswas significantly larger than wild-type Synechococcus

(1.76 versus 1.61 mm), which resulted in a significantincrease in cell volume (3.89 versus 3.08 mm3; Fig. 3E;Supplemental Table S2). Overall, these results indicatea role for hydrocarbons in limiting cell size and en-suring normal cell division.

Hydrocarbons Are Essential for Optimal Cell Growth

Strains were then cultured under continuous moderatelight (40 mmol photons m22 s21) to determine whether alack of hydrocarbons in the membrane affected growth.Due to the difference in cell size between wild type andhydrocarbon-deficient mutants, which affects the opticalproperties of the culture (Fig. 4, A–D), growth was mea-sured both by cell counting and by optical density. Theincrease in cell number during exponential growth wasapproximately 4-fold higher in wild-type Synechocystiscultures, compared to DFAR (Fig. 4A). Moreover, photo-bleaching increased in DFAR cells after 2 d growth, asmeasured by the amount of chlorophyll per cell (Fig. 4E).This suggests that cell damage was occurring during thistime. The enlarged phenotype of DFAR was maintainedover this growth period (Fig. 4G).Wild-type Synechococcusalso demonstrated a statistically significant 1.4-fold increasein cell number during exponential growth compared toDOls (Fig. 4B), although photobleaching was not observed(Fig. 4F). Under moderate light, when starting with anequal amount of culture as determined by opticaldensity, growth of the wild type was 2.2-fold fasterthan DFAR (Supplemental Fig. S5A). Growth of wild-typeSynechococcuswas 1.5-fold faster than DOls (SupplementalFig. S5B). Thedifference in growth rates betweenwild-typeSynechocystis and Synechococcus and the hydrocarbon-deficient mutants was similar at a higher light intensityof 120mmol photonsm22 s21 (Supplemental Fig. S5, C andD). Overall, these results demonstrate the importance ofhydrocarbons for optimal cell growth.

The Absence of Hydrocarbons Has Minor Effects onPhotosynthetic Performance

Other cellular traits were then examined to determinewhether these could affect cell growth. The maximumphotosynthetic rate, as measured by oxygen evolutionper unit of chlorophyll, was not reduced in the DFARand DOls mutants (Fig. 5, A and B). An increase inrespiration was observed in DFAR cells, with a 2-foldhigher rate observed compared to wild type (Fig. 5C).In algae, respiration increases with cellular size (Tangand Peters, 1995), and our data suggest that the samerelationship may occur in cyanobacteria. Despite theincreased respiratory rate, growth of DFAR was stillimpaired under light/dark cycles (Supplemental Fig.S6). However, respiration was similar between wildtype and DOls (Fig. 5D). Photoinhibition was alsocomparable between wild-type and hydrocarbon-deficient strains (Fig. 5, E and F).

The absorbance profile and emission spectra of thephotosynthetic and light harvesting complexes were

Figure 2. Hydrocarbons accumulate within cyanobacterial membranes.A and B, Detection of CP47 (A) and SbtA (B) in purified plasma andthylakoid membrane fractions (replicates 1–3). Small amounts of CP47were consistently detected in the purified plasma membrane fractions.Five micrograms of protein was loaded with antibodies diluted 1:2000.PM, Plasma membrane; TM, thylakoid membrane. C, Percentage of hy-drocarbons as total lipids in Synechocystis total, thylakoid and plasmamembranes, and Synechocystis total membrane samples. Results arefrom three biological replicates. Mean 6 SD is indicated. Statistical sig-nificance was determined by a Student’s t test.

1930 Plant Physiol. Vol. 172, 2016

Lea-Smith et al.



















then examined. Absorbance was slightly reduced in bothhydrocarbon-deficient mutants in the 400 to 550 nm range(Supplemental Fig. S7), the portion of the spectra corre-sponding to carotenoid and chlorophyll absorption. How-ever, the carotenoid/chlorophyll ratio was not significantlydifferent between strains (Supplemental Table S5), suggest-ing that the altered absorbance profile of the hydrocarbon-deficient mutants could be due to differences in lightscattering,whichhaveagreater effect at shorterwavelengthsin the spectrum. Analysis of the hydrocarbon-deficientmutants via 77K fluorescence emission spectra showed mi-nor but consistent differences in energy transfer efficiencyfromphycobilisomes to the reaction centers of photosystemsinDFAR andDOls, a blue shift in the peak between 680 and700nm in DFAR, indicative of increased uncoupling ofphycobilisomes from photosystems (Supplemental Fig. S8,A and B), and an altered photosystem II to photosystem Iratio (Supplemental Fig. S8, C andD).Given that the oxygenevolution rates of the hydrocarbon-deficient strains aresimilar to wild type, the cumulative effect of these changeson photosynthetic efficiency must be minor. Overall, theseresults suggest that differences in cell size and division maybe the major factors in the impaired growth observed inhydrocarbon-deficient mutants.

Hydrocarbons May Induce Membrane Flexibility byAccumulating within the Lipid Bilayer

Molecular dynamics simulations have become aninvaluable technique used to investigate the nanoscale

organization of lipid membranes (Marrink et al., 2009;Vattulainen and Rog, 2011), particularly in complexmembrane systems (Ingólfsson et al., 2014; Mannaet al., 2014). In order to understand how hydrocarbonscould affect membrane properties, a novel symmetricalmembrane model system was simulated based on thepseudoatomistic Martini force field, with an approxi-mately 4:1 mapping of heavy atoms to coarse-grainedparticles (Supplemental Fig. S9; López et al., 2013). Thepresentmodelused16different lipid types corresponding tothe four major groups present in cyanobacteria: phosphati-dylglycerol (PG), monogalactosyl-diacylglycerol (MGDG),digalactosyl-diacylglycerol (DGDG), and sulfoquinovosyl-diacylglycerol (SQDG), in a ratio as experimentally deter-mined in Synechocystis (Supplemental Table S6; Sheng et al.,2011). The system contained a total of 2400 lipids, resultingina largemembrane slabwithdimensionsof approximately21 3 27 nm. The hydrocarbon heptadecane was addedrandomly to the solvent of the equilibratedmembranes after2 ms and observed to enter the bilayer within the first 50 psof simulation due to its hydrophobicity. Heptadecane be-came fully incorporated within;20 ns, remained solvatedwithin the membrane for the full 5 ms of simulation, andwas localized between the two monolayers, alongside thelipid tails at the center of the bilayer (Fig. 6, A–D).

In symmetrically modeled membranes where no flip-flopping of individual lipids across leaflets occurs, likethe one studied here, a flat lamellar bilayer would beexpected. This was the case in the absence of alkanes, inwhich a stable, noncurved membrane was observed

Figure 3. Hydrocarbon-deficient mutants have increased cell size and division defects. A, Bright-field images of wild-typeSynechocystis (left) and DFAR (right) cells. Scale bars correspond to 5 mm. B and C, Cell volume of Synechocystis strainsquantified via measuring the diameter of cells from confocal microscopy images (B) and particle counting measurements (C). D,Bright-field images of wild-type Synechococcus (left) and DOls (right) cells. Scale bars correspond to 5 mm. E, Cell volume ofSynechococcus strains quantified via measuring the width and length of cells from confocal microscopy images. B, C, and E,Mean6 SD is indicated. Statistical significancewas determined by a Student’s t test. F, Percentage of single versus actively dividingSynechocystis cells. Statistical significance was determined by a two-way x2 test.












(Fig. 6A). Addition of hydrocarbons led to their spon-taneous insertion and clustering within the bilayer core,with a concomitant increase in membrane thicknessfrom ;3.27 nm to ;3.95 nm, irrespective of concen-tration. The overall lipid lateral diffusion coefficients inall systems were within experimentally reported ranges(Kaňa, 2013). Pools of clustered hydrocarbon moleculeswere associated with a reduction in lipid chain orderand packing efficiency, particularly at $5% mol/molhydrocarbon concentrations (Supplemental Fig. S10).Moreover, increasing amounts of hydrocarbon dis-solved within the bilayer center, which led to localizedswelling on one side of the membrane, around the sitesof hydrocarbon accumulation, as visually evident in thecross sections (Fig. 6, B–D). The swelling settled in onedirection or another, and this direction did not changeduring the simulation, presumably due to the stochastic

distribution of solubilized hydrocarbons within themembrane. This is consistent with neutron diffractionstudies, which indicated alkane incorporation andswelling of dioleoyl lecithin bilayers (White et al., 1981).The accumulation of hydrocarbons thus increased theflexibility of the membrane and induced localizedswelling. It should also be noted that the use of an alter-native lipid parameter set developed for the membranesof Thermosynechococcus vulcanus and Spinacia oleracea (vanEerden et al., 2015) similarly induced swelling and dis-order in our bilayer model in the presence of alkanes.

The level of swelling observed at $7.5% mol/molhydrocarbons due to the presence of a large hydrocar-bon pool eventually destabilized the membrane,resulting in a phase transition to a nonlamellar bilayer.In a macroscopic system and/or under conditions offixed simulation volume, the membrane swelling and

Figure 4. Hydrocarbons are essential for optimalgrowth of Synechocystis and Synechococcus. A toD, Growth of Synechocystis (A and C) and Syn-echococcus (B and D) under 40 mmol photonsm22 s21 light. An equal number of cells, approxi-mately 5 3 106 for Synechocystis strains and 1 3107 for Synechococcus strains, was added to eachculture. This corresponded to an OD750nm equal to0.07 6 0.001, 0.127 6 0.009, and 0.06 6 0.002for wild-type Synechocystis, DFAR, and ΔFAR:comp (C), respectively, and 0.076 6 0.003 and0.105 6 0.008 for wild-type Synechococcus andDOls (D), respectively. The growth rate constantsfor wild-type Synechocystis, DFAR, and ΔFAR:compwere 1.246 0.203 106, 3.086 0.133 105

(P = 0.0169), and 1.52 6 0.14 3 106 cells h21,respectively, and for wild-type Synechococcusand DOls were 1.98 6 0.02 3 106 and 1.42 60.04 3 106 cells h21 (P = 0.0009), respectively.E and F, The amount of chlorophyll per cell (inattomol) in Synechocystis and Synechococcusstrains, respectively. G, Cell diameter of Synecho-cystis strains. Results are from three biologicalreplicates. Errors bars indicate SD. Asterisks indi-cate significant differences betweenwild-type andhydrocarbon-deficient samples (Student’s paired ttest: P , 0.05).


Lea-Smith et al.




lipid disorder would be expected to result in inductionof significant bilayer curvature. Typically, membranecurvature depends upon induced asymmetry of onemonolayer compared to another (McMahon and Gallop,2005). Local clustering of nonbilayer forming lipids couldalso lead to curvature. MGDG is one such lipid, whereasPG, DGDG, and SQDG favor flat lamellar phases(Shipley et al., 1973; Tilcock, 1986), and local MGDGenrichment could hinder the formation of completelamellar bilayer phases, even in combination withother thylakoid lipids (Murphy, 1982).

Synechococcus Hydrocarbon-Deficient MutantsDemonstrate Reduced Membrane Curvature

To assess the effects of hydrocarbon deficiency on mem-brane conformation in Synechocystis and Synechococcus, weused thin-section electron microscopy. Electron micro-graphsof thewild-type andhydrocarbon-deficientmutantssuggested that the thylakoid membranes are more planarin the mutants, although this effect could only be properlyquantified and verified in Synechococcus, due to its moreregular thylakoid membrane layout and its elongated cell

shape. In thin-section images from Synechococcus, we se-lected cells that appeared circular in profile: In these cases,we could be sure that the thin-section cut across the cellperpendicular to the long axis, since any other sectionwould be noncircular (Supplemental Fig. S11). In the cir-cular sections, the thylakoidmembranes appear as an arrayof roughly parallel membrane sacs, each spanning the gapbetween a pair of poorly defined bodies close to the plasmamembrane termed the “thylakoid centers” (Kunkel, 1982;Stengel et al., 2012). Typically, each thin section showedtwo to four thylakoid centers distributed around the cellperimeter, with the thylakoid membrane sacs extendingbetween them.

To derive a quantitative measure of membrane cur-vature, we traced the membrane between two thylakoidcenters and measured its length and also measured thestraight-line distance between the thylakoid centers (Fig.6E). The ratio of these two measures reflects the curva-ture of the membrane. We measured the curvature ofover 100 membrane segments from each strain. Therewas no significant difference between the means of thewild type and DOls internode distances. On average,thylakoid membranes in wild-type cells were found tobe more curved than those of DOls (Fig. 6F). The mean

Figure 5. Photosynthetic rates and photoinhibitionare similar between wild-type and hydrocar-bon-deficient mutants. A and B, Oxygen evo-lution was measured at different light intensitiesin Synechocystis (A) and Synechococcus (B) toexamine photosynthesis. The maximum photo-synthetic rate of wild-type Synechocystis,DFAR,and DFAR:comp was 0.311 6 0.025, 0.329 60.024, and 0.29960.028mmolO2 nmolChl21 h21,respectively, and of wild-type Synechococcus andDOls was 0.283 6 0.03 and 0.303 6 0.018 mmolO2 nmol Chl21 h21, respectively. C and D, Respi-ration was determined by measuring oxygenconsumption following each light period inSynechocystis (C) and Synechococcus (D). Theaverage oxygen consumption rate followingdark periods after 95 mmol photons m22 s21 ofwild-type Synechocystis, DFAR, and DFAR:comp was 0.041 6 0.008, 0.083 6 0.006, and0.043 6 0.01 mmol O2 nmol Chl21 h21, re-spectively, and of wild-type Synechococcus andDOls was 0.027 6 0.005 and 0.027 6 0.007mmol O2 nmol Chl21 h21, respectively. E and F,Photoinhibition was determined by measuringoxygen evolution at a light intensity of 2000 mmolphotons m22 s21 in Synechocystis (E) and3000 mmol photons m22 s21 in Synechococcus(F). All results are from six to nine separate bi-ological replicates. Errors bars indicate SD. As-terisks indicate significant differences betweenwild-type and hydrocarbon-deficient samples(Student’s paired t test: P , 0.05).






length ratio was 1.09 6 0.06 in wild type versus 1.06 60.07 inDOls, with the relatively high standard deviationsreflecting a range of membrane curvatures in both thewild type and DOls (Fig. 6F; Supplemental Fig. S12).Nevertheless, the difference in the means is highly sig-nificant, with a P value of 0.00007 from a Student’s t test.

DISCUSSION

Here, we have shown a role for hydrocarbons intwo morphologically different cyanobacterial species.While both hydrocarbon-deficient mutants display in-creased cell size, division defects, and reduced growth,a more severe phenotype was observed in DFAR cells(Figs. 3 and 4). Spherical cells have a larger fraction ofhighly curved membranes than rod-shaped cells. In thecase of cyanobacteria, greater membrane flexibilitywould be required in order to incorporate multiplethylakoid membranes and to divide efficiently. High-resolution inelastic neutron scattering experiments ofSynechocystis cells demonstrated dynamic flexibilitywithin thylakoid membranes that differed betweenlight and dark periods, suggesting that, if hydrocarbons

affect curvature, these compounds may also have a rolein other cellular functions (Stingaciu et al., 2016). Whilethe division dynamics of cyanobacteria are poorlyunderstood, in the spherical bacterium Staphylococcusaureus, cells divide by first forming a septum, leadingto development of two daughter cells connected via anarrow peripheral ring, followed by an abrupt sepa-ration event (Zhou et al., 2015). This form of divisioninduces high stress on cellular components and is de-pendent on extreme curvature in membranes. A sim-ilar division event in Synechocystis and other sphericalcyanobacteria requires the induction of membranecurvature not only in the cytoplasmic membrane butalso in the thylakoid membranes, in order that theseare efficiently distributed between daughter cells. Bycontrast, rod-shaped cells divide by first increasing involume and length, followed by formation of a sep-tum in themiddle of the extended cell and subsequentseparation (Wu and Errington, 2011). This form of celldivision would require less induction of membranecurvature in the cytoplasmic membrane and thyla-koid membranes and would be necessary at only oneend of the cell. Interestingly, in the DOls strain, hy-drocarbons were more important for efficient daughter

Figure 6. Hydrocarbons disrupt membrane order by integrating into the lipid bilayer. A to D, Modeling of cyanobacterialmembranes containing (A) 0, (B) 2.5, (C) 5, and (D) 7.5 mol heptadecane/mol total lipids in the bilayer. Hydrocarbons are shownas red van derWaals spheres. Lipids are shown in stick representation and colored as follows: lipid head group rings are shown inmagenta, phosphate beads in tan, sulfate beads in yellow, diglycerol beads in pink, and lipid tails in cyan. Snapshots show thedirection of swelling associated with alkane accumulation that the membranes settled into, which was stochastic. E, Electronmicrograph of a transverse section of Synechococcus, illustrating measurement of the curvature index, given by the ratio of thelength of the membrane section (yellow line) and the shortest distance between the ends of the membrane section (cyan line). F,Comparison of lines with curvature index derived frommembrane measurements. Mean6 SD is indicated. Statistical significancewas determined by a two-tailed t test. The distance of the internode measurements was similar between strains (wild type: 6166147 nm, DOls: 582 6 142 nm; two-tailed t test: P = 0.1, a = 0.05).


Lea-Smith et al.




cell separation than division, as shown by the formationof chains of cells (Fig. 3D).Although the DOls mutant showed significantly less

thylakoid membrane curvature on average than wild-type Synechococcus, examples of membrane curvaturecould be observed in this strain (Fig. 6; SupplementalFig. S11), despite the natural tendency of lipid bilayersto adopt a flat shape (Graham and Kozlov, 2010).Moreover, since simulations indicated that they inte-grate into the middle of the bilayer (Fig. 6, B–D), hy-drocarbons would be unable to orient the direction ofcurvature, suggesting that their major role may be toinduce the required flexibility in membranes. Therefore,hydrocarbons cannot be the only factor determiningmembrane curvature: Other factors must contribute toboth the direction and maintenance of curvature. Inaddition it was observed that after successive rounds ofsubculturing, typically six to eight, that the size differ-ence between the hydrocarbon-deficient mutants andwild-type strains was reduced and Synechococcus DOlscells no longer formed chains of cells. That suggests thatother factors in the cell were compensating for the loss ofhydrocarbons.AnArabidopsis (Arabidopsis thaliana) protein, CURT1A,

has been shown to induce membrane curvature in chlo-roplast membranes (Armbruster et al., 2013). A homolo-gous protein in Synechocystis, CurT, has recently beenshown to have a similar role in thylakoid membranes(Heinz et al., 2016). Deletion of CurT resulted in a re-duction in growth and extreme differences in thylakoidmembrane organization,with the thylakoids appearing tocross the cytoplasm and not converging on the “thylakoidcenters.” In contrast to DFAR, cell size was not affected,although photosynthesis was reduced. The DcurT strainalso displayed disassociated phycobilisomes, similarto what was observed in DFAR and DOls (SupplementalFig. S8, A and B). This strongly suggests that the de-gree of membrane curvature is essential for optimalphycobilisome:photosystem interaction and may alsoinfluence contact of other soluble proteinswithmembrane-bound components. Therefore, it is possible that, if hy-drocarbons do alter membrane curvature, then this isaugmented and orientated by CurT. In DcurT, the thyla-koid membranes were still highly curved, indicatingthat other factors are involved in inducing membranecurvature (Heinz et al., 2016). Homologs of CurT arepresent in the majority of sequenced cyanobacterialstrains (Supplemental Table S1). Notable exceptionsinclude Gloeobacter species, which lack thylakoidmembranes (Rippka et al., 1974; Rexroth et al., 2011;Saw et al., 2013) and therefore may not require orien-tation of membrane curvature or may regulate it byother means. In other bacterial species, this includesturgor pressure or force applied via cytoskeletalcomponents (Cabeen et al., 2009). The glycolipidMGDG may also help stabilize this curvature, givenits tendency to favor nonlamellar phases (Shipleyet al., 1973; Murphy, 1982; Tilcock, 1986). Otheras-yet-unidentified factors may also contribute tomembrane curvature.

Hydrocarbons may also have additional functions incells not identified in this study, such as modulatingmembrane permeability (Valentine and Reddy, 2015).The use of planar lipid bilayers as model systems hasdemonstrated that the addition of hexadecane increasesmembrane thickness and reduces membrane permea-bility (Dilger and Benz, 1985). Therefore, the increase incell size may be due to a combination of factors: differ-ences in osmotic pressure due to reduced membranepermeability, outward physical pressure on the cellapplied by a series of less-curved thylakoid mem-branes, and division impairment, which would resultin hydrocarbon-deficient strains being larger thanwild type before cell separation. However, in the caseof Synechococcus cells, it is interesting that an increasein size was only observed along the long axis of thecell, where outward physical pressure applied by less-curved thylakoid membranes would be expected tohave the greatest effect.

CONCLUSION

Given that maintaining optimal growth and cell di-vision is important in all ecosystems (Raven, 1998), therole of hydrocarbons in supporting optimal growththrough potentially inducing membrane flexibility andreducing membrane permeability may be sufficient toexplain the strong evolutionary pressure to retain hy-drocarbon biosynthesis in cyanobacteria. It may alsoexplain why similar hydrocarbons are produced bysome microalgae species. An additional advantage isthat unlike phospholipids or proteins, hydrocarbons donot contain either phosphorus or nitrogen, which arelimited inmany environments, notably in the open oceanwhere Synechococcus and Prochlorococcus speciesdominate (Flombaum et al., 2013). Moreover, thenonreactive properties of hydrocarbons make themresistant to oxidative damage (Valentine and Reddy,2015), which is a major issue in cyanobacteria due toconstant electron production from photosynthesis andrespiration (Lea-Smith et al., 2016a). Hydrocarbon-induced membrane curvature may therefore repre-sent a unique, low-risk, and efficient system of inducingflexibility and reducing permeability in one of the mostbiologically important and ancient membrane systemson the planet.

MATERIALS AND METHODS

Bioinformatics

Protein BLAST comparisons (Altschul et al., 1990) were performed usinginferred protein sequences for Synechocystis sll0209 (FAR), sll0208 (FAD), andslr0483 (CurT) and Synechococcus Syn7002_A1173 (Ols; WP_012306795) withthe completed cyanobacterial genomes listed in the National Center for Bio-technology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/genome/browse/) and Biller et al. (2014). For FASTA BLAST comparisons ofOls, only matches across the majority of the gene (.90%) were included, dueto the conservation of many domains in polyketide synthase proteins. ProteinBLAST comparisons of FAR, FAD, and Ols were also performed against thebacterial and eukaryotic sequences in the NCBI database in order to confirmthat these proteins are cyanobacterial specific.









http://www.ncbi.nlm.nih.gov/genome/browse/

http://www.ncbi.nlm.nih.gov/genome/browse/


Bacterial Strains, Media, and Growth Conditions

Synechocystis and Synechococcus strains were routinely cultured in BG11mediumwith 10 mM sodium bicarbonate (Castenholz, 1988) at 30°C and grownunder moderate light (30–40 mmol photons m22 s21) with shaking at 120 rpmunless otherwise indicated. HEPES and vitamin B12were added to Synechococcuscultures to a final concentration of 10 mM and 4 mg/L, respectively. Ten milli-molar sodium bicarbonate was also added to Synechococcus cultures every 2 d.Fifteen grams/liter of agar was used for preparation of solid media and sup-plementedwith 30 to 100mg/mLof kanamycin or 5%Suc (w/v)when necessary.Cultures incubated at 120 mmol photons m22 s21 were bubbled with air to pre-vent carbon limitation.

Plasmid Construction

All primers are listed in Supplemental Table S7. Polymerase chain reactionswere performed by standard procedures using Phusion high-fidelity DNApolymerase (New England Biolabs). The genome sequence of Synechocystis andSynechococcus (Kaneko et al., 1996) was consulted via Cyanobase (http://genome.kazusa.or.jp/cyanobase) for primer design. Gene deletion of Sll0209was performed by amplifying a 1750-bp fragment spanning Sll0209 andflanking regions using primers Sll0209for and Sll0209rev, followed by insertioninto the XbaI/SphI sites of pUC19. The aph gene conferring kanamycin resis-tance was excised from pUC4K (Vieira andMessing, 1982) and inserted into theHincII site in the middle of the fragment to generate pSll0209. Gene deletion ofols (SYNPCC7002_A1173) was performed by amplifying a 1922-bp fragment inthe 59 region using primers Olsfor and Olsrev, followed by insertion into theEcoRI/SalI sites of pUC19. The aph gene was inserted into the blunt endedBamHI in the middle of the fragment to generate pOls.

Gene deletion of phaAB was performed by amplifying a 1069-bp fragmentupstream of phaA using primers PhaABleftfor and PhaABleftrev and a 1087-bpfragment downstream of phaB using primers PhaABrightfor and PhaABrightrev,followed by insertion of the respective fragments into theXbaI/BamHI and SacI/EcoRI sites of pUC19 to generate pPhaAB-1. The BamHI-digested npt1/sacRBcassette from pUM24Cm (Ried and Collmer, 1987) was inserted into the BamHIsite between the upstream and downstream fragments to generate pPhaAB-2. Togenerate the plasmid for complementation (pSll0209comp) ofDSll0209, the entireSll0209 gene plus 295 bp of upstream region and 263 bp of downstream regionwas amplified using primer pairs Sll0209compfor and Sll0209comprev andinserted into the BamHI/SacI sites of pPhaAB-1.

Generation of Mutant Strains

Generation of marked mutants was conducted according to Lea-Smith et al.(2013, 2016b). Approximately 1 mg of plasmids pSll0209, pOls, and pPhaAB-2was mixed with Synechocystis or Synechococcus cells for 6 h in liquid medium,followed by incubation on BG11 agar plates for approximately 24 h. An addi-tional 3 mL of agar containing kanamycin was added to the surface of the platefollowed by further incubation for approximately 1 to 2 weeks. Transformantswere subcultured to allow segregation of mutant alleles. In the case of thehydrocarbon-deficient mutants, this was performed by streaking the strains onBG11 agar plates containing 30mg/mL of kanamycin, followed by a subsequentrestreak on a BG11 agar plate containing 100 mg/mL of kanamycin. Typically,segregatedmutants were obtainedwithin 2 weeks. This is in contrast to a recentreport, in which hydrocarbon-deficient Synechocystis mutants were onlyobtained after approximately 6 months, most likely due to these strains beingsegregated on BG11 agar plates containing a maximum of 40 mg/mL of kana-mycin (Berla et al., 2015). Repeated streaking over a 6-month period could alsoresult in selection of numerous secondary mutations. Given that a com-plemented strain was not generated or examined in the Berla et al. (2015) study,it is therefore impossible to determine whether the phenotype observed wascaused by deletion of hydrocarbons or secondary mutations. Due to this factorand the difference in time in generating mutants, a direct comparison betweenthe results reported by Berla et al. (2015) and this study is difficult due to theinstability of the hydrocarbon-deficient strains.

Segregation was confirmed by PCR using primers Sll0209for/Sll0209rev,Olsfor/Olsrev, or Phafor/Pharev, which flank the inserted region (SupplementalFig. S1). Generation of unmarked mutants was carried out according to Xu et al.(2004) and Lea-Smith et al. (2013, 2016b). To remove the npt1/sacRB cassette andinsert the Sll0209 complementation cassette, the phaAB-marked knockout wastransformed with 1 mg of the markerless pSll0209comp construct. Followingincubation in BG11 liquid medium for 4 d and agar plates containing Suc for a

further 1 to 2 weeks, transformants were patched on kanamycin and Suc plates.Suc-resistant, kanamycin-sensitive strains containing the unmarked deletionwere confirmed by PCR using primers flanking the insert region (SupplementalFig. S1B). TheDSll0209mutantwas generated in theDPhaAB:Sll0209 backgroundin order to produce the complement strain.

The DOls mutant could not be complemented due to the large size of thegene (8163 bp). Therefore, wild-type Synechococcus and DOls were sequencedusing the Illumina MiSeq personal sequencer and mapped to the Synechococcusgenome. Apart from the expected deletion in ols, only a single point mutationin DOls, leading to a silent mutation, was observed when compared to thewild type.

For characterization, mutant strains were subcultured in liquid medium nomore than two times and streaked on solidmediumamaximumof six times, dueto the instability of themutants. After this period, the size difference between thehydrocarbon-deficient mutants and wild-type strains was reduced, suggestingthat another factor in the cell was compensating for the loss of hydrocarbons.Strains could not be prepared as glycerol stocks, since this also resulted in achange in the phenotype. After this period of subculturing, fresh mutants wereconstructed for analysis.

Extraction and Analysis of Total Hydrocarbons

All chemicals were purchased from Sigma chemicals. For extraction of totalhydrocarbons, 1.5 mL of dichloromethane was added to pelleted dried cellsin glass vials, and hydrocarbons were extracted and analyzed by gaschromatography-mass spectrometry (GC-MS) according to Lea-Smith et al.(2015). Hydrocarbons and lipids were extracted from Synechocystis thylakoid,cytoplasmic, and total membrane fractions, and Synechococcus total membranefractions based on the method in Davey et al. (2008) where 1 mL (3 mL for totalmembrane fractions) of chilled (220°C) solvent (methanol:chloroform:water,2.5:1:1) was added to the membrane fraction tube, vortexed, and left in ice withoccasional shaking. After 30 min, tubes were centrifuged (16,000g, 2 min, 4°C).The supernatant was removed and placed in a chilled tube on ice. Theremaining pellet was re-extracted with 0.5 mL (1.5 mL for total membranefractions) chilled (220°C) methanol:chloroform 1:1 for 30 min. After centri-fuging as described previously, the supernatants were combined in a 2mL tube.The organic chloroform phase was separated from the aqueous phase byadding 250 mL (750 mL) chilled water and extracted into a new glass 2 mL GCsample vial. The chloroform phase was dried (GeneVac EZ-2; SP Scientific) andresuspended in 200 mL heptane. The extracts were stored at 280°C beforeanalysis of total alkanes and lipids (fatty acid methyl esters [FAMEs]). Fornegative controls, extraction blanks were carried out without cyanobacteria (nosignificant amounts of hydrocarbons were detected) and positive controlsconsisted of adding 1 mg/mL standard alkane mix (Sigma C8-C20 alkane mix)to a blank extraction procedure.

Purification of Membrane Fractions

Plasma and thylakoid membranes were isolated using a combination of Sucdensity centrifugation followedbyaqueous two-phase partitioning according toNorling et al. (1998) and Huang et al. (2002). Synechocystis cells were grown at30°C under 50 mmol photons m22 s21 of white light in BG11 medium withbubbling air. All steps were carried out at 4°C unless otherwise stated. Twoliters of cells harvested at OD750nm = 0.9 to 1.0 were resuspended in buffer A(20 mM potassium phosphate, pH 7.8) and broken with glass beads. Unbrokencells and debris was pelleted by centrifugation at 3000g for 10 min. The su-pernatant was centrifuged at 103,000g for 30 min to pellet total membranes.Total membranes were made up to a concentration of 42% Suc by the additionof solid Suc and placed onto a discontinuous Suc gradient comprising of 3 mLlayers of 50% (w/w), 42% (w/wwith totalmembranes), 40% (w/w), 38% (w/w),35% (w/w), 30% (w/w), and 10% (w/w) Suc in buffer A and centrifuged at125,000g for 15 h. The fraction between 38% and 42%was collected, diluted withbuffer A, and centrifuged at 125,000g for 45 min to pellet membranes. Pelletedmembranes were homogenized in buffer B (5 mM potassium phosphate, pH 7.8,0.25 M Suc) to a weight of 3.75 g and applied to a 6.25 g polymer mixture of 5.8%Dextran T-500 and 5.8% polyethylene glycol 3350 in buffer B. The partitioningsystem was gently inverted 35 times at 4°C and centrifuged at 1000g to facilitatephase separation. Pure thylakoid membranes were obtained from the lowerphase after five further partitionings in the 5.8% polymer mixture. Pure plasmamembraneswere obtained from the ninth upper phase after three partitionings inthe initial polymermixture of 5.8%, three further partitionings in 6.2%, and a finalthree partitionings in 6.4%. Purified plasma and thylakoid membranes were


Lea-Smith et al.



http://genome.kazusa.or.jp/cyanobase

http://genome.kazusa.or.jp/cyanobase






diluted in buffer B and pelleted by centrifugation at 125,000g for 1 h and ho-mogenized in a minimal volume of the same buffer.

Identification and Quantification of Hydrocarbons in theMembrane Fractions

Hydrocarbons in the heptane extract were identified by GC-MS (ThermoScientific TraceGC1310-ISQLTSingleQuadrupleEIMS,A1-1310Autosampler)with a Phenomenex Zebron ZB-5MSi Capillary GC Column (30 m3 0.25mm30.25 mm). The injection volume was 1 mL with a 10:1 split ratio with an injectortemperature of 300°C, using helium as a carrier gas at a constant flow of 1.0 mLmin21. The following gradient was used: initial oven temperature 70°C, 2 min;76°C, 1 min; 250°C at 6°C min21; and 330°C at 50°C min21. The transfer linetemperature was 250°C. Themass spectrometry conditions in the positivemodewere ion source, 250°C; mass range 45 to 650 D; and scan time of 0.35 s.Heptadecane and nonadecene were identified by coretention with stan-dards and NIST mass spectral search libraries (National Institute of Stan-dards and Technology NIST v2.0); 8-heptadecene was identified using theNIST library alone. Heptadecane and nonadecene were quantified usingstandard curves derived from peak areas of heptadecane and nonadecenestandards; 8-heptadecene was quantified using peak areas derived fromheptadecane standards (0.06–31 mg/mL).

Identification and Quantification of Total Lipids in theMembrane Fractions

The total lipid content of the heptane extract was converted to FAMEs asdescribed by Davey et al. (2014). The FAMEs were separated and identifiedusing GC-MS as described in the membrane alkane analysis section but with a35:1 split injection ratio, injector temperature of 230°C, helium at a constant flowrate of 1.2 mLmin21, and with the following gradient: initial oven temperature,60°C for 2 min; 150°C at 15°C min21; and 230°C at 3.4°C min21. The detectortemperature was 250°C with a scan time of 0.174 s. FAMEs were identified bycoelution with a FAME standards and NIST libraries and were quantified andsummed using standard curves derived from C16:0 methyl esters.

Modeling of Synechocystis Membranes

The in silico cyanobacterial membrane lipid compositions were based on theexperimental lipid extractions and characterization of Synechocystis by Shenget al. (2011). Four major classes of cyanobacterial lipids were used: PG, MGDG,DGDG, and SQDG, with various acyl tails differing in length and degree ofsaturation. These lipids contain a palmitic (16:1D9) tail at the sn-2 position andanother acyl tail of variable length and saturation at the sn-1 position (Murataet al., 1992). The composition of the lipid tails in the in silico membranes wasadapted to coarse-grained resolution, i.e. an approximately 4:1 mapping ofheavy atoms to coarse-grained particles, using the Martini force field (Lópezet al., 2013). The structure of the lipid head groups and representative tails in-cluded in the model are compared to their coarse-grained topologies inSupplemental Figure S9, where the mapping of the Martini bead types areshown and labeled. Standard bonded parameters were used. The compositionsdetermined by Sheng et al. (2011) are shown in Supplemental Table S6, andcompared to the number of lipids used in our model membranes to reproduceas closely as possible this composition. The hydrocarbon heptadecane wasadded in varying quantities to study the effects of this compound onmembraneproperties.

A total of 2400 lipids were used to build symmetric bilayers, consisting of16 different lipid types. The system was solvated with 11,323 water beads,corresponding to ;45,000 waters, ensuring that the bilayer was well hydrated.Hydrated sodium counterion particles were added to neutralize the charges ofPG and SQDG lipids. Heptadecane was added randomly to the solvent in theequilibrated membranes after 2 ms. The amounts used were 2.5, 5.0, and 7.5 mol%, corresponding to 60, 120, and 180 alkane molecules, respectively. The initialunit cell dimensions of all membrane systems were 21.0 3 27.53 9.0 nm in thex, y, and z directions.

Simulation Details

The molecular dynamics simulations were performed using the GROMACS4.5.5 MD package. The Martini lipid force field was used (Marrink et al., 2007)due to its proven performance in describing complex lipid membrane

properties. Initially, a system containing 200 randomly placed 18:1 DGDGmolecules surrounded by solvent was simulated for 200 ns, yielding a pre-equilibrated bilayer. The lipid types were then converted at random to yielda membrane with the appropriate composition (Supplemental Table S6), usingin-house code. Following minimization and a further 200 ns equilibration, thecoordinates of this bilayer systemwere thenmultiplied in the x- and y-dimensionsto produce the full 2400 lipid bilayer. A 1 ms equilibration simulation followed.Steepest descent was used for minimization, and a 40 fs time step was used to-getherwith the leap-frog algorithmduring simulations. Lennard-Jones (excludingscaled 1–4) interactionswere smoothly switched off between 0.9 and 1.2 nmusinga force switch. Electrostatic interactions were calculated using a shifted potentialwith a cutoff of 1.2 nm, with a distance-dependent dielectric constant of 15. Theneighbor list of 1.4 nm was updated every 10 steps. The isothermal-isobaric en-semble was used. The pressure (1 bar) and temperature (316 K) coupling pa-rameters were set to 5 ps semi-isotropically, and 10 ps, respectively(Berendsen et al., 1984). All systems were simulated for 5 ms.

Electron Microscopy

Synechococcus cultures were grown to OD750nm = ;0.3 and harvested bycentrifugation (3000g; 10 min), fixed, and embedded according to the protocoldescribed in Nürnberg et al. (2014) using potassium permanganate as addi-tional fixative. Thin sections were cut with a glass knife at a Reichert Ultracut Emicrotome and collected on uncoated, 300 mesh copper grids. High contrast wasobtained by poststaining with saturated aqueous uranyl acetate and lead citrate(Reynolds, 1963) for 4 min each. The grids were examined in a JOEL JEM-1230transmission electron microscope at an accelerating potential of 80 kV.

Curvature Measurements

In transverse sections, the thylakoid membranes, three to five membranesthick, of wild-type Synechococcus andDOls cells appear to emanate from three tofour well-spaced nodes on the edge of the cell, like pages that fan out from thespine of a book that is lying open. The “spine” of this “book” can be imagined torun longitudinally along the cell frompole to pole. Spline curveswere hand-fittedto individualmembrane layers as far as they could be traced by eye from node tonode in ImageJ. The ratio of the length of the curved line drawn to its Feret di-ameter (maximum caliper distance, i.e. the straight line distance between startand end points of the line in most cases) was taken to be a measure of its cur-vature. The 124 membranes from nine wild-type cells and 102 membranes from12 DOls cells were analyzed. Statistical tests were performed in Matlab.

Confocal Fluorescence Microscopy

For confocalmicroscopy,midlogarithmicphase cellswere spottedontoBG111% (w/v) agar plates and visualized with a Leica laser-scanning confocal mi-croscope SP5 using a363 oil-immersion objective (Leica HCX PL APO lambdablue 63.031.40 OIL UV). Chlorophyll a fluorescence was detected by using anexcitation wavelength of 488 nm and an emission range from 670 to 720 nm.Images were captured with a pinhole of 95.5 mm, which corresponds to anoptical section thickness of 0.8 mm and by 43 line averaging. Analyses wereperformed with ImageJ 1.47i (http://imagej.nih.gov/ij) and Origin. The cellvolumewas determined from themean diameter for Synechocystis cells assuminga sphere, and from the mean diameter and width of Synechococcus cells by as-suming an ellipsoidal shape. A Student’s t test was used for comparison of cellvolumes between strains with P, 0.05 being considered statistically significant.

Cell Counting

Numbers of cells per unit of volume were measured by counting the cellsdirectly using a Beckman Coulter 2Z particle counter. Measurements wereperformed by diluting 20 to 100mL of cells in 10mLofmeasuring buffer. The celldiameter of Synechocystis cells was directly measured using the same instru-ment. Cell volume was calculated from these measurements. Due to the rodshape of Synechococcus, the cell size of this bacterium could not be determinedusing this device.

Cell Division

Asemiautomated counting of cellswasused todetermine the number of cellsthat were in the process of division by segmenting the image based on







http://imagej.nih.gov/ij


fluorescence intensity and cell size. A frequency table of the number of cellsobserved to be above andbelow the size threshold (interpreted to be dividing anddivided, respectively) was generated (Supplemental Table S4). A two-way x2 testyielded a significant result, showing that the proportions of dividing versus di-vided cells is not independent of the strain and therefore suggests that there aremore dividing cells in DFAR than in the wild type with statistical significance.

Measurements of Cell Growth

Growth rate constants as determined by cell counting were calculated duringearly exponential phase (0–46 h and 0–44 h, respectively, for Synechocystis andSynechococcus strains cultured under 40mmol photonsm22 s21 light). Growth rateconstants as determined by optical density were calculated during early expo-nential phase (0–40 h and 0–90 h, respectively, for Synechocystis strains culturedunder 40 and 120mmol photonsm22 s21 light; 0–26 h and 0–90 h, respectively, forSynechococcus strains cultured under 40 and 120 mmol photons m22 s21 light; and18–78 h for Synechocystis strains cultured under 12 h light [40 mmol photons m22

s21/12 h dark cycles]). A Student’s paired t test was used for comparison ofgrowth between strains, P , 0.05 being considered statistically significant.

Chlorophyll Measurements

The amount of chlorophyll in Synechocystis samples was measured bysubtracting the 750 nm optical density (OD) value from the 680 nm OD valueand multiplying the total by 10.854, according to Lea-Smith et al. (2013). Todetermine the correlation between OD values versus the chlorophyll concen-tration of Synechococcus, a range of samples was measured at 750 nm and680 nm, in addition tomeasuring the chlorophyll concentration according to themethod of Porra et al. (1989). A strong correlation (r2 = 0.9983) was observed(Supplemental Fig. S13). The amount of chlorophyll in Synechococcus sampleswas then measured by subtracting the 750 nm OD value from the 680 nm ODvalue and multiplying the total by 12.959.

Photosynthesis, Photoinhibition, andRespiration Measurements

Photoinhibition, photosynthesis, and respiration were determined according toLea-Smith et al. (2014). Photosynthetic O2 evolution rates and O2 depletion rates(respiration) were determined on cell cultures at OD750nm = ;0.5 (;2.3 nmol Chlml21 in Synechocystis or ;4 nmol Chl ml21 in Synechococcus) using an oxygenelectrode system (Hansatech) maintained at 30°C. DFAR cell cultures were col-lected at an OD750nm = ;0.4 and concentrated to an OD750nm = ;0.5 prior toanalysis. Following dark equilibration (10 min), O2 exchange rates were recordedfor 10min at increasing light intensities (10, 20, 50, 95, 240, 450, 950, and 2000 mmolphotons m22 s21), using Realite MR16+C 24°, 12 V, 50 W C13 white LED lamps(Deltech), which have a spectra similar to sunlight. Each light period was followedimmediately by 10min indarkness to calculate the respiration rates. The respirationrate following illumination at each light intensity periodwas subtracted to estimatethe real rate of photosynthetic O2 evolution. To measure photoinhibition, cell cul-tures of OD750nm =;0.2 (;1 nmol Chl ml21 in Synechocystis or;1.3 nmol Chl ml21 in Synechococcus)werefirst dark equilibrated (10min), and the rate ofO2 evolutionwas recorded for 50 min at a light intensity of 2000 mmol photons m22 s21 inSynechocystis and 3000 mmol photons m22 s21 in Synechococcus. All measurementswere standardized to the initial rate. A Student’s paired t test was used for allcomparisons, P , 0.05 being considered statistically significant.

77 K Fluorescence

Weperformed 77Kfluorescencemeasurements on cells harvestedduring theexponential growth phase at an OD750nm = ;0.3, diluted to a final chlorophyllconcentration of 5 mM and placed into glass sample tubes. After dark adapta-tion at room temperature for approximately 10 min, samples were then snap-frozen in liquid nitrogen. We recorded 77 K fluorescence emission spectra by aPerkin Elmer LS55 fluorescence spectrometer from 620 to 800 nm with either600 nm (phycobilisome excitation) or 435 nm (chlorophyll excitation).

Absorbance Measurements

Absorbance measurements on whole cells were performed according toLea-Smith et al. (2014). Cultures were harvested during the exponential

growth phase at an OD750nm = ;0.4. Cultures were placed in a 4 mL fluores-cence cuvette (1 cm path length) and positioned in front of the entrance port ofan integrating sphere. A light source sent light via an input fiber into thecuvette containing the sample and the light leaving the sample in the forwarddirection was collected by the integrating sphere. The extinction spectra wererecorded using a USB4000-UV-VIS Ocean Optics spectrometer connected tothe integrating sphere with an output fiber optic and interfaced to a computer.The cuvettes containing the samples were positioned at different distances(0 mm and 5 mm) from the entrance port of the sphere, and the absorbancespectrumwas obtained via the SpectraSuite Spectroscopy operating software.The nominal absorption spectrum was then calculated using the equationaccording to Merzlyak and Naqvi (2000).

Carotenoid Quantification

HPLC was performed to analyze carotenoid/chlorophyll a ratios. Pigmentswere extracted from freeze-dried samples (triplicates) by three subsequent ex-traction steps in ice-cold methanol. After addition of methanol, the sampleswere vortexed vigorously, incubated on ice for 10 min, and centrifuged for10 min at 4°C at 14,000 rpm. The supernatants of all three extraction steps werecombined and filtered using 13 mm, 0.22 mm PTFE syringe disc filters. Twohundred microliters of each sample was loaded onto a Dionex HPLC system,which was equipped with a LiChrospher 100 RP-18 (5 mm) reverse-phase col-umn (Merck 1.50943.0001). The flow rate was set at 1 mL/min and the mobilephase composed of two eluents (A, 0–12 min; B, 12–23 min; A: 87% acetonitrile,10% methanol, 3% 0.1 M Tris buffer pH 8; B: 80% methanol, 20% hexane). Pig-ments were detected spectrophotometrically at 447 nm, and absorbance spectradata were collected from 350 to 750 nm. The relative quantity of each pigmentresolved was determined by integration of the area under the 447 nm peak(mAU3min). Pigments were identified using published absorbance spectradata (Mohamed and Vermaas, 2004).

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Generation of mutant strains in Synechocystis andSynechococcus.

Supplemental Figure S2. Chromatograms showing separation of hydro-carbons from whole cells.

Supplemental Figure S3. Chromatograms showing separation of hydro-carbons from membrane fractions.

Supplemental Figure S4. Bright-field confocal images of Synechocystis andSynechococcus strains.

Supplemental Figure S5. Growth of Synechocystis and Synechococcus undermoderate and high light.

Supplemental Figure S6. Growth of Synechocystis under moderate light/dark cycles.

Supplemental Figure S7. Absorbance profiles of Synechocystis and Syne-chococcus strains.

Supplemental Figure S8. 77K fluorescence of Synechocystis and Synecho-coccus strains.

Supplemental Figure S9. CG topologies of representative lipids of themembrane.

Supplemental Figure S10. Membrane partial density, lipid acyl chain or-der parameter, and organization in differing hydrocarbon contents.

Supplemental Figure S11. Electron microscopy images of Synechococcuscells.

Supplemental Figure S12. Length ratios (curvatures) of membranes fromSynechococcus 7002 strains.

Supplemental Figure S13. Correlation between the absorbance at 680 nmand 750 nm, and amounts of chlorophyll measured following methanolextraction.

Supplemental Table S1. Conservation of hydrocarbon biosynthetic path-way proteins and CurT in sequenced cyanobacteria strains.


Lea-Smith et al.



















Supplemental Table S2. Cell size as determined via fluorescence microscopy.

Supplemental Table S3. Cell size as determined via particle counting mea-surements.

Supplemental Table S4. Cell counts of single and actively dividing Syn-echocystis cells.

Supplemental Table S5. Carotenoid/chlorophyll ratios in cyanobacterialstrains.

Supplemental Table S6. Lipid composition of cyanobacterial membranes.

Supplemental Table S7. Sequence of primers used in this study.

ACKNOWLEDGMENTS

We thank Pietro Cicuta (Department of Physics, University of Cambridge),Gabriele Kaminski Schierle (Department of Chemical Engineering and Biotech-nology, University of Cambridge) and James Locke, Arijit Das, and BrunoMartins (Sainsbury Laboratory, Cambridge) for useful discussion.

Received August 1, 2016; accepted October 3, 2016; published October 5, 2016.

LITERATURE CITED

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic localalignment search tool. J Mol Biol 215: 403–410

Armbruster U, Labs M, Pribil M, Viola S, Xu W, Scharfenberg M, HertleAP, Rojahn U, Jensen PE, Rappaport F, et al (2013) Arabidopsis CUR-VATURE THYLAKOID1 proteins modify thylakoid architecture by in-ducing membrane curvature. Plant Cell 25: 2661–2678

Berendsen HJC, Postma JPM, van Gunsteren WF, Dinola A, Haak JR(1984) Molecular dynamics with coupling to an external bath. J ChemPhys 81: 3684–3690

Berla BM, Saha R, Maranas CD, Pakrasi HB (2015) Cyanobacterial alkanesmodulate photosynthetic cyclic electron flow to assist growth under coldstress. Sci Rep 5: 14894

Biller SJ, Berube PM, Berta-Thompson JW, Kelly L, Roggensack SE,Awad L, Roache-Johnson KH, Ding H, Giovannoni SJ, Rocap G, et al(2014) Genomes of diverse isolates of the marine cyanobacterium Pro-chlorococcus. Sci Data 1: 140034

Bourdenx B, Bernard A, Domergue F, Pascal S, Léger A, Roby D, Pervent M,Vile D, Haslam RP, Napier JA, et al (2011) Overexpression of ArabidopsisECERIFERUM1 promotes wax very-long-chain alkane biosynthesis and in-fluences plant response to biotic and abiotic stresses. Plant Physiol 156: 29–45

Cabeen MT, Charbon G, Vollmer W, Born P, Ausmees N, Weibel DB,Jacobs-Wagner C (2009) Bacterial cell curvature through mechanicalcontrol of cell growth. EMBO J 28: 1208–1219

Castenholz RW (1988) Culturing methods for cyanobacteria. MethodsEnzymol 167: 68–93

Choi YJ, Lee SY (2013) Microbial production of short-chain alkanes. Nature502: 571–574

Coates RC, Podell S, Korobeynikov A, Lapidus A, Pevzner P, ShermanDH, Allen EE, Gerwick L, Gerwick WH (2014) Characterization ofcyanobacterial hydrocarbon composition and distribution of biosyn-thetic pathways. PLoS One 9: e85140

Davey MP, Burrell MM, Woodward FI, Quick WP (2008) Population-specific metabolic phenotypes of Arabidopsis lyrata ssp. petraea. NewPhytol 177: 380–388

Davey MP, Horst I, Duong GH, Tomsett EV, Litvinenko ACP, Howe CJ,Smith AG (2014) Triacylglyceride production and autophagous re-sponses in Chlamydomonas reinhardtii depend on resource allocation andcarbon source. Eukaryot Cell 13: 392–400

Dilger JP, Benz R (1985) Optical and electrical properties of thin monooleinlipid bilayers. J Membr Biol 85: 181–189

Flombaum P, Gallegos JL, Gordillo RA, Rincón J, Zabala LL, Jiao N, KarlDM, Li WK, Lomas MW, Veneziano D, et al (2013) Present and futureglobal distributions of the marine Cyanobacteria Prochlorococcus andSynechococcus. Proc Natl Acad Sci USA 110: 9824–9829

Galloway JN, Dentener FJ, Capone DG, Boyer EW, Howarth RW, SeitzingerSP, Asner GP, Cleveland CC, Green PA, Holland EA, et al (2004) Nitrogencycles: past, present, and future. Biogeochemistry 70: 153–226

Graham TR, Kozlov MM (2010) Interplay of proteins and lipids in gen-erating membrane curvature. Curr Opin Cell Biol 22: 430–436

Heinz S, Rast A, Shao L, Gutu A, Gugel IL, Heyno E, Labs M, Rengstl B,Viola S, Nowaczyk MM, et al (2016) Thylakoid membrane architecturein Synechocystis depends on CurT, a homolog of the granal CURVATURETHYLAKOID1 proteins. Plant Cell 28: 2238–2260

Howard RW, Blomquist GJ (2005) Ecological, behavioral, and biochemicalaspects of insect hydrocarbons. Annu Rev Entomol 50: 371–393

Howard TP, Middelhaufe S, Moore K, Edner C, Kolak DM, Taylor GN,Parker DA, Lee R, Smirnoff N, Aves SJ, et al (2013) Synthesis of cus-tomized petroleum-replica fuel molecules by targeted modification offree fatty acid pools in Escherichia coli. Proc Natl Acad Sci USA 110:7636–7641

Howe CJ, Barbrook AC, Nisbet RER, Lockhart PJ, Larkum AWD (2008)The origin of plastids. Philos Trans R Soc Lond B Biol Sci 363: 2675–2685

Huang F, Parmryd I, Nilsson F, Persson AL, Pakrasi HB, Andersson B,Norling B (2002) Proteomics of Synechocystis sp. strain PCC 6803:identification of plasma membrane proteins. Mol Cell Proteomics 1: 956–966

Ingólfsson HI, Melo MN, van Eerden FJ, Arnarez C, Lopez CA, WassenaarTA, Periole X, de Vries AH, Tieleman DP, Marrink SJ (2014) Lipid orga-nization of the plasma membrane. J Am Chem Soc 136: 14554–14559

Kaňa R (2013) Mobility of photosynthetic proteins. Photosynth Res 116:465–479

Kaneko T, Sato S, Kotani H, Tanaka A, Asamizu E, Nakamura Y, MiyajimaN, Hirosawa M, Sugiura M, Sasamoto S, et al (1996) Sequence analysis ofthe genome of the unicellular cyanobacterium Synechocystis sp. strainPCC6803. II. Sequence determination of the entire genome and assignmentof potential protein-coding regions (supplement). DNA Res 3: 185–209

Kosma DK, Bourdenx B, Bernard A, Parsons EP, Lü S, Joubès J, Jenks MA(2009) The impact of water deficiency on leaf cuticle lipids of Arabidopsis.Plant Physiol 151: 1918–1929

Kunkel DD (1982) Thylakoid centers: structures associated with the cya-nobacterial photosynthetic membrane system. Arch Microbiol 133: 97–99

Lea-Smith DJ, Biller SJ, Davey MP, Cotton CA, Perez Sepulveda BM,Turchyn AV, Scanlan DJ, Smith AG, Chisholm SW, Howe CJ (2015)Contribution of cyanobacterial alkane production to the ocean hydro-carbon cycle. Proc Natl Acad Sci USA 112: 13591–13596

Lea-Smith DJ, Bombelli P, Dennis JS, Scott SA, Smith AG, Howe CJ (2014)Phycobilisome-deficient strains of Synechocystis sp. PCC 6803 have reducedsize and require carbon-limiting conditions to exhibit enhanced productivity.Plant Physiol 165: 705–714

Lea-Smith DJ, Bombelli P, Vasudevan R, Howe CJ (2016a) Photosynthetic,respiratory and extracellular electron transport pathways in cyanobac-teria. Biochim Biophys Acta 1857: 247–255

Lea-Smith DJ, Ross N, Zori M, Bendall DS, Dennis JS, Scott SA, SmithAG, Howe CJ (2013) Thylakoid terminal oxidases are essential for thecyanobacterium Synechocystis sp. PCC 6803 to survive rapidly changinglight intensities. Plant Physiol 162: 484–495

Lea-Smith DJ, Vasudevan R, Howe CJ (2016b) Generation of marked andmarkerless mutants in model cyanobacterial species. J Vis Exp doi/10.3791/54001

Li N, Chang WC, Warui DM, Booker SJ, Krebs C, Bollinger JM Jr (2012)Evidence for only oxygenative cleavage of aldehydes to alk(a/e)nes andformate by cyanobacterial aldehyde decarbonylases. Biochemistry 51:7908–7916

López CA, Sovova Z, van Eerden FJ, de Vries AH, Marrink SJ (2013)Martini force field parameters for glycolipids. J Chem Theory Comput 9:1694–1708

Manna M, Róg T, Vattulainen I (2014) The challenges of understandingglycolipid functions: an open outlook based on molecular simulations.Biochim Biophys Acta 1841: 1130–1145

Marrink SJ, de Vries AH, Tieleman DP (2009) Lipids on the move: sim-ulations of membrane pores, domains, stalks and curves. Biochim Bio-phys Acta 1788: 149–168

Marrink SJ, Risselada HJ, Yefimov S, Tieleman DP, de Vries AH (2007)The MARTINI force field: coarse grained model for biomolecular sim-ulations. J Phys Chem B 111: 7812–7824

McMahon HT, Gallop JL (2005) Membrane curvature and mechanisms ofdynamic cell membrane remodelling. Nature 438: 590–596

Mendez-Perez D, Begemann MB, Pfleger BF (2011) Modular synthase-encoding gene involved in a-olefin biosynthesis in Synechococcus sp.strain PCC 7002. Appl Environ Microbiol 77: 4264–4267











Merzlyak MN, Naqvi KR (2000) On recording the true absorption spec-trum and the scattering spectrum of a turbid sample: application to cellsuspensions of the cyanobacterium Anabaena variabilis. J PhotochemPhotobiol B 58: 123–129

Mohamed HE, Vermaas W (2004) Slr1293 in Synechocystis sp. strain PCC6803 is the C-39,49 desaturase (CrtD) involved in myxoxanthophyll bi-osynthesis. J Bacteriol 186: 5621–5628

Murata N, Wada H, Gombos Z (1992) Modes of fatty-acid desaturation incyanobacteria. Plant Cell Physiol 33: 933–941

Murphy DJ (1982) The importance of non-planar bilayer regions in photo-synthetic membranes and their stabilization by galactolipids. FEBS Lett150: 19–26

Norling B, Zak E, Andersson B, Pakrasi H (1998) 2D-isolation of pureplasma and thylakoid membranes from the cyanobacterium Synecho-cystis sp. PCC 6803. FEBS Lett 436: 189–192

Nürnberg DJ, Mariscal V, Parker J, Mastroianni G, Flores E, MullineauxCW (2014) Branching and intercellular communication in the Section Vcyanobacterium Mastigocladus laminosus, a complex multicellular pro-karyote. Mol Microbiol 91: 935–949

Porra RJ, Thompson WA, Kriedemann PE (1989) Determination of accurateextinction coefficients and simultaneous-equations for assaying chlorophyllsa and b extracted with 4 different solvents: verification of the concentration ofchlorophyll standards by atomic-absorption spectroscopy. Biochim BiophysActa 975: 384–394

Raven JA (1998) The twelfth Tansley Lecture. Small is beautiful: the pico-phytoplankton. Funct Ecol 12: 503–513

Rexroth S, Mullineaux CW, Ellinger D, Sendtko E, Rögner M, Koenig F(2011) The plasma membrane of the cyanobacterium Gloeobacter violaceuscontains segregated bioenergetic domains. Plant Cell 23: 2379–2390

Reynolds ES (1963) The use of lead citrate at high pH as an electron-opaquestain in electron microscopy. J Cell Biol 17: 208–212

Ried JL, Collmer A (1987) An nptI-sacB-sacR cartridge for constructingdirected, unmarked mutations in gram-negative bacteria by markerexchange-eviction mutagenesis. Gene 57: 239–246

Rippka R, Waterbury J, Cohenbazire G (1974) Cyanobacterium whichlacks thylakoids. Arch Microbiol 100: 419–436

Saw JHW, Schatz M, Brown MV, Kunkel DD, Foster JS, Shick H,Christensen S, Hou S, Wan X, Donachie SP (2013) Cultivation andcomplete genome sequencing of Gloeobacter kilaueensis sp. nov., from alava cave in Kilauea Caldera, Hawai’i. PLoS One 8: e76376

Schirmer A, Rude MA, Li X, Popova E, del Cardayre SB (2010) Microbialbiosynthesis of alkanes. Science 329: 559–562

Sheng J, Vannela R, Rittmann BE (2011) Evaluation of methods to extract andquantify lipids from Synechocystis PCC 6803. Bioresour Technol 102: 1697–1703

Shipley GG, Green JP, Nichols BW (1973) The phase behavior of mo-nogalactosyl, digalactosyl, and sulphoquinovosyl diglycerides. BiochimBiophys Acta 311: 531–544

Sorigué D, Légeret B, Cuiné S, Morales P, Mirabella B, Guédeney G, Li-BeissonY, Jetter R, Peltier G, Beisson F (2016) Microalgae synthesize hydrocarbonsfrom long-chain fatty acids via a light-dependent pathway. Plant Physiol 171:2393–2405

Stengel A, Gügel IL, Hilger D, Rengstl B, Jung H, Nickelsen J (2012) Initialsteps of photosystem II de novo assembly and preloading with manganesetake place in biogenesis centers in Synechocystis. Plant Cell 24: 660–675

Stingaciu LR, O’Neill H, Liberton M, Urban VS, Pakrasi HB, Ohl M(2016) Revealing the dynamics of thylakoid membranes in living cya-nobacterial cells. Sci Rep 6: 19627

Tan X, Yao L, Gao Q, Wang W, Qi F, Lu X (2011) Photosynthesis drivenconversion of carbon dioxide to fatty alcohols and hydrocarbons in cy-anobacteria. Metab Eng 13: 169–176

Tang EPY, Peters RH (1995) The allometry of algal respiration. J PlanktonRes 17: 303–315

Tilcock CP (1986) Lipid polymorphism. Chem Phys Lipids 40: 109–125Valentine DL, Reddy CM (2015) Latent hydrocarbons from cyanobacteria.

Proc Natl Acad Sci USA 112: 13434–13435van Eerden FJ, de Jong DH, de Vries AH, Wassenaar TA, Marrink SJ (2015)

Characterization of thylakoid lipid membranes from cyanobacteria and higherplants bymolecular dynamics simulations. Biochim Biophys Acta 1848: 1319–1330

Vattulainen I, Rog T (2011) Lipid simulations: a perspective on lipids inaction. Cold Spring Harb Perspect Biol 3: a004655

Vieira J, Messing J (1982) The pUC plasmids, an M13mp7-derived systemfor insertion mutagenesis and sequencing with synthetic universalprimers. Gene 19: 259–268

White SH, King GI, Cain JE (1981) Location of hexane in lipid bilayersdetermined by neutron diffraction. Nature 290: 161–163

Wu LJ, Errington J (2011) Nucleoid occlusion and bacterial cell division.Nat Rev Microbiol 10: 8–12

Xu H, Vavilin D, Funk C, Vermaas W (2004) Multiple deletions of smallCab-like proteins in the cyanobacterium Synechocystis sp. PCC 6803:consequences for pigment biosynthesis and accumulation. J Biol Chem279: 27971–27979

Zhou X, Halladin DK, Rojas ER, Koslover EF, Lee TK, Huang KC, Theriot JA(2015) Bacterial division. Mechanical crack propagation drives milliseconddaughter cell separation in Staphylococcus aureus. Science 348: 574–578

Zwirglmaier K, Jardillier L, Ostrowski M, Mazard S, Garczarek L, VaulotD, Not F, Massana R, Ulloa O, Scanlan DJ (2008) Global phylogeographyof marine Synechococcus and Prochlorococcus reveals a distinct partitioningof lineages among oceanic biomes. Environ Microbiol 10: 147–161


Lea-Smith et al.



Hydrocarbons Are Essential for Optimal Cell Size, · a role for hydrocarbons in limiting cell size...

Documents

Transcript of Hydrocarbons Are Essential for Optimal Cell Size, · a role for hydrocarbons in limiting cell size...